KR20100072770A - Hot data management based on hit counter from data servers in parallelism - Google Patents

Abstract

PURPOSE: A hot data management method based on access number of times collected form data servers in dispersion are provide to perform monitoring and resolving hot data through effective dispersion of a load even if request of many users is generated at predetermined time in asymmetry storage system, thereby providing stable data service. CONSTITUTION: An access number about data which each data server of an asymmetric storage system stores is monitored. Each data server maintains access number information per data in recent state. Each data server transmits the access number information in predetermined cycle to a meta data server. According to crystallization of the metadata server, each data server performs deletion of a duplicate or copy(S971).

Description

Hot Data Management Based on Hit Counter from Data Servers in Parallelism

The present invention relates to an asymmetric storage system, and more particularly, to a hot data management method based on the number of accesses that can prevent the load of the data server due to the hot data by efficiently managing the hot data in the asymmetric storage system.

The present invention is derived from the research conducted as part of the IT growth engine technology development project of the Ministry of Knowledge Economy and the Ministry of Information and Communication Research and Development. [Task management number: 2007-S-016-02] ].

Most mass storage systems adopt asymmetrical structures in which metadata and data are processed separately to ensure high scalability, metadata is managed by the metadata server, and data is managed by the data server. Here, metadata means location information of a data server that stores actual data of a file.

The data server that stores and manages each data provides the actual data stored on disk by the user's request through the network. Depending on the disk performance of the data server or the transmission performance of the network, the data server can serve one data server. There is a limit to performance.

For example, when a large video service such as User Created Contents (UCC) is provided, a large number of accesses to a specific video file in a certain period of time, many read requests to the data server that stores and manages the data. However, since the data service provides only the highest performance of the disk or network, a failure (for example, a break in the video) occurs not only for the additional data service but also for the existing video service of the user.

In an asymmetric storage system, when a large number of users encounter a intensive read request for a specific file over a period of time (hereinafter referred to as hot data), the physical performance of the data server storing and managing the data of the specific file (ie , The performance of disks and networks) can not provide a smooth data service. To solve this problem, if you try to detect and resolve the hot data by the metadata access count of a single metadata server instead of the data server, you cannot track the number of file read requests, which is the actual load of the data. In addition, since the access count value must be updated every time the metadata is accessed, a lot of load is generated.

In addition, even if the hot data is once hot data, the hot data is not continuously maintained after a certain time. Without considering the characteristics of these hot data, additionally duplicated data wastes storage to solve the hot data.

The present invention is to solve the above problems, to provide a hot data management method based on the number of accesses that can prevent the load concentration of the data server due to hot data by efficiently managing the hot data in an asymmetric storage system There is an object of the present invention.

Another object of the present invention is to provide a hot data management method based on the number of accesses that can provide a stable data service by detecting and solving hot data through efficient distribution of load even if a large number of user requests occur at a given time in an asymmetric storage system. To provide.

Another object of the present invention is to collect the number of data access in each data server and send it to the metadata server in a certain period of time to detect, resolve and track the hot data, so that even if hot data occurs, the user can smoothly provide a read data service to the user. To provide a hot data management method based on the number of accesses.

Another object of the present invention is to track and manage the list of hot data already resolved through replication, and the number of allocated data when the number of times is not hot data. To provide a way.

In order to achieve the above object, the present invention monitors the number of times of access to data stored by each data server of the asymmetric storage system, and keeps the access number information for each data up to date; A server transmitting the access count information to a metadata server at predetermined intervals, and each data server performing copying or deleting a copy of data according to the determination of the metadata server. Provides a hot data management method.

According to another aspect of the invention, the method comprising the steps of constructing a hot data management table including an access count field for each data, collecting access count information for data stored in each data server from one or more data servers; Updating the hot data management table according to the access count information, checking the hot data management table at predetermined intervals to determine whether the hot data is present, and replicating the data file determined as hot data to a new data server. And deleting a copy of the data file determined to be no longer hot data.

According to another aspect of the invention, a plurality of data server to maintain the number of times of data access to the data stored and managed by the plurality of data server for a predetermined period, the plurality of data server for the predetermined period for the stored data Transmitting the number of data accesses to a management server; collecting and storing the number of data accesses transmitted by the management server; and number of accesses of each data server by the management server at a predetermined period and every cycle time. Recognizing data exceeding a predetermined threshold as the corresponding hot data, and additionally replicating the data of the hot file to one or more data servers of the plurality of data.

According to the present invention, even if a large number of user requests occur at a given time in an asymmetric storage system, by detecting and solving hot data through efficient load distribution, a stable data service can be provided.

In addition, the number of data accesses are collected from each data server and sent to the metadata server within a certain period of time to detect, resolve, and track the hot data, thereby smoothly providing a read data service to the user even if the hot data occurs.

In addition, by tracking and managing the hot data list already solved through replication, the number of allocated data is counted when it is no longer hot data, thereby preventing waste of storage.

Hereinafter, with reference to the accompanying drawings will be described in detail a preferred embodiment of the present invention.

1 is a configuration diagram of an asymmetric storage system to which embodiments of the present invention are applied.

Referring to FIG. 1, an asymmetric storage system to which embodiments of the present invention are applied includes n user file systems 110-1 through 110-n, a metadata server 120, and m data servers 130. -1 to 130-m).

User file systems 110-1 through 110-n receive a file related user's request. At this time, the user file systems 110-1 to 110-n request metadata from the metadata server 120 according to a user's request, and the data servers 130-1 to 130-m request the metadata. Requires real data

The metadata server 120 managing location information on the actual data of the file checks the validity of a request of a specific user file system among user file systems 110-1 through 110-n, and requests metadata (ie Location information of file data) is transmitted to the specific user file system via a network.

The data servers 130-1 to 130-m managing the actual data of the file deliver data of the disk to the specific user file system according to the request of the specific user file system.

In order to collect information regarding the occurrence of the read load, the data servers 130-1 to 130-m maintain the number of data accesses for the user's data read request with the configuration as shown in FIG. 2.

The data access number entries 210 to 212 include a disc identifier field and a data identifier field for distinguishing data, an access count field for recording access times, a hash list field and a top list field, and the user access request. Thus, a hash function 201 for retrieving fast data access count entries, and a hash table 202 with MAX hash heads are used.

Hash function 201 has the data identifier of the user request, and gets the result of dividing the data identifier by MAX (data identifier% MAX). This hash result is the value of one entry in the hash table 202.

If a hash result value equal to the result of dividing another data identifier by MAX is obtained, the data access number entries are prev as shown in the data access number entry 210 and the data access number entry 211 shown in FIG. 2. It is chained through a hash list field containing the entry and next entry information.

In order to maintain the ranking of the number of accesses of each data, the data access number entries 210 to 212 are used to display the prev and next entry information of each data access number entry, starting from the top list head. Form a connection using the top list field including a.

In the example shown in FIG. 2, the data access number entry 211 is located after the top list head, and then the data access number entry 212 and the data access number entry 210 are sequentially concatenated. It is preferable that the data access number entry 211 close to the higher priority, that is, the data access number entry for a file having many access times.

3 is a flowchart illustrating a method of processing data access times according to an embodiment of the present invention. Hereinafter, the data server is any one of the data servers 130-1 to 130-m of FIG. 1.

As shown in FIG. 3, when the data server is started, the data server initializes the hash table (S301) and waits for a user request (S302). In this case, when a user request for reading or deleting is received (S310), the following procedure is performed to update the number of data accesses.

That is, the data server substitutes the data identifier into the hash function to obtain a hash result value (S311). Subsequently, the data server obtains an entry of a hash table, that is, a hash list head as the hash result value (S312), and then inserts a next entry into a temporary entry (S313).

The data server determines whether the temporary entry and the hash list head are the same (S320). If the result is the same, the data server determines whether the disk identifier and the data identifier of the user's request are the same (S330). If the disc identifier and the data identifier are the same as the result of the determination, the data server determines whether the user request is a delete request or a read request (S340).

If it is determined that the user request is a deletion request, the data server removes the corresponding data access number entry from the hash list (S341), and also removes the corresponding data access number entry from the top list (S342). The data server removes the corresponding data access number entry (S343).

If the user request is a read request, the data server increments the access count of the corresponding data access number entry once (S351) and updates the top list of the corresponding data access count entry (S352). ).

If the temporary entry and the hash list head are the same as the determination result at operation S320, the data server creates a new data access number entry, inserts a disk identifier and a data identifier into the generated data access number entry, and accesses the generated data. The number of accesses of the count entry is initialized to '1' (S361). Subsequently, the data server puts a hash list of the generated data access number entries into the hash list head (S362), and puts a top list of the generated data access number entries into the last portion of the top list head (S363).

If the disc identifier and the data identifier are the same as the result of the determination in operation S330, the data server inserts the next entry into the temporary data access number entry in operation S371 and proceeds to determination in operation S320.

In summary, when a data read or delete request is received, the data server inserts a data identifier into the hash function to obtain a result, and then checks for an entry in the hash table, that is, a data access count entry with the data identifier in the hash list head. do. If the next entry is a read request, the access count field is incremented by one, the top list of entries is updated, and if there is no entry, a new data access count entry is created, initialized, and put into the hash list head. If the entry exists and the delete request is made, the corresponding data access count entry is removed.

4 is a flowchart illustrating a method of changing a rank of data access times according to an embodiment of the present invention.

Referring to FIG. 4, the data server fetches the previous entry from the top list of the corresponding data access number entry having the increased data access count (S401), and puts the previous entry into the temporary entry (S402).

The data server determines whether the temporary entry is the same as the top list head (S410). If the determination result is not the same, the data server determines whether the access number of the temporary entry is greater than or equal to the access number of the data access number entry (S420).

As a result of the determination, when a temporary entry is found that is equal to or greater than the number of accesses, the corresponding data access number entry is designated next to the top list of the temporary entry (exactly, the data access number entry currently inserted into the temporary entry) (S423). In addition, a temporary entry is specified before the data access count entry. Through this process, it is possible to change the rank between entries of data access times.

If the determination result in the determination process (S420) does not find an entry with the same or greater number of times, the data server retrieves the previous entry from the top list of the temporary entries (S421), and inserts the obtained previous entry into the temporary entry (S422). . In this way, entries with higher priority, that is, more accesses, are searched.

If an entry with the same or higher number of accesses than the corresponding data access entry is found up to the head of the top list, then the data access entry is the entry with the highest priority, so the data server is currently in the top list of data access entry entries. The contents are removed (S431), and the top list of data access number entries is designated as the next entry of the top list head (S432).

5 is a flowchart illustrating a method of processing data access times according to an embodiment of the present invention.

Each data server collects the number of data accesses (S510), determines whether a predetermined period has been reached (S520), and when the set period is reached, checks the maximum number of transmissions (S530), and transmits the number to the metadata server. The access count information is transmitted (S540). Thereafter, the access count entry is initialized (S550).

On the other hand, the step of checking the maximum number of transmissions (S530) need not be performed every time, of course.

6 is a diagram illustrating a configuration of a hot data management table by a metadata server that receives access count information according to an embodiment of the present invention.

As shown in FIG. 6, the hot data management table to be stored in a database (DB) included in the metadata server includes a file (inode) identifier field 601, a recent access time field 602, and an access count field. 603, and an additional replica field 604.

The file identifier field 601 is a value for distinguishing a file and is an inode identifier in a virtual file system (VFS).

The recent access time field 602 maintains recent time information of updating the access count value of the file data received from each data server.

The additional copy presence field 604 is for tracking and managing hot data, and indicates additional copy to solve the hot data.

Reference numeral 620 denotes a configuration of the access count field 603. The fields 621-1 to 621-60, min [0], min [1] to min [59], represent file access count values every minute based on the last_min (621) time. 1 to 631-24), hour [0] to hour [23], represent a file access count value every hour based on the last_hour 631 time. Days [0] through day [364], which are fields 641-1 through 641-365, represent file access counts per day based on the last_day 641 time.

7 is a flowchart illustrating a method of storing data access count information according to an embodiment of the present invention. That is, FIG. 7 illustrates a process of storing data access count information (for example, disk identifier, data identifier, and access count) received from each data server in the hot data management table. Hereinafter, the metadata server is the same server as the metadata server 120 of FIG. 1, and the data server is a data server of any one of the data servers 130-1 to 130-m of FIG. 1.

Referring to FIG. 7, the metadata server receives a data access count value from the data server (S701), and obtains a file identifier to which a disk identifier and a data identifier belong (S702).

Then, it is determined whether a file identifier corresponding to the disk identifier and the data identifier exists (S710), and if not present, it is regarded as a deleted file, and the process returns to the reception process (S701) to obtain other data access count information.

If it is determined that the file identifier exists, the metadata server determines whether a data access number entry having the file identifier exists in a predetermined hot data management table (S720).

If there is a data access number entry having the file identifier as a result of the determination, the metadata server fetches the access count field of the hot data management table (S721), and specific fields of the access count field, for example, a min [0] value. The hour [0] value and the day [0] value are increased by the data access count value (S722), the latest access time is updated to the current time (S723), and the access count field is updated.

If it is determined in operation S720 that the data access number entry having the file identifier does not exist, the metadata server adds a new data access number entry having the file identifier to the hot data management table in operation S731. . The metadata server checks all fields of the access count field of the new data access count entry, that is, min [0] to min [63], hour [0] to hour [23] and day [0] to day [364]. Initialize to '0' (S732). Subsequently, the metadata server sets last_min, last_hour, and last_day in the access count field of the new data access count entry as the current time (S733), and proceeds to increase the field (S722).

8 is a flowchart illustrating a method of updating an access count field according to an embodiment of the present invention. That is, FIG. 8 shows the number of access fields (ie, min [0] to min [59], hour [0] to hour [23], and day [0] to day [364]) of the hot data management table. It shows the process of updating on the basis of. Hereinafter, the metadata server is the same server as the metadata server 120 of FIG. 1.

Referring to FIG. 8, the metadata server compares last_day of the current time and the access count field to determine whether last_day has elapsed for more than one day (S810).

If it is determined that more than one day has passed, the metadata server shifts the day [0] to day [364] of the access count field to the right by the elapsed day (S811), and from day [0] to day [elapsed day- 1] field is initialized to '0' (S812). The metadata server initializes min [0] to min [59] and hour [0] to hour [23] of the access count field to '0' (S813), and sets last_min, last_hour and last_day as the current time. Update (S814).

If it is determined that the determination process (S810) has not elapsed more than one day, the metadata server compares the current time and last_hour in the access count field to determine whether last_hour has elapsed by at least one hour (S820).

If more than one hour has passed as a result of determination in the determination process (S820), the metadata server shifts hour [0] to hour [23] in the access count field to the right by an elapsed time (S821), hour [0] to hour [elapsed time-1] are initialized to '0' (S822). Then, min [0] to min [59] in the access count field are initialized to '0' (S823), and last_min and last_hour are updated to the current time (S824).

If it is determined that the determination process (S820) has not elapsed more than an hour, the metadata server compares last_min of the current time and the access count field to determine whether last_min has elapsed for more than 1 minute (S830). If not more than 1 minute has passed, the metadata server terminates the update of the number of access fields.

If more than 1 minute has passed as a result of the determination in the determination process (S830), the metadata server shifts min [0] to min [59] in the access count field to the right by the number of minutes (S831), and min [ 0] to minute [elapsed minute-1] are initialized to '0' (S832). The metadata server updates last_min with the current time (S833).

9 is a flowchart illustrating a method of detecting and duplicating hot data according to an embodiment of the present invention. Hereinafter, the metadata server is the same server as the metadata server 120 of FIG. 1.

Referring to FIG. 9, the metadata server obtains a check execution cycle from the hot data management table (S901), and sets a check period and an access count threshold from the hot data management table to detect hot data. Bring (S902). The metadata server retrieves all entries whose latest access time is within the inspection period from the current time from the hot data management table (S903), checks the imported entries one by one (S904), and checks whether it is hot data (S910). . If the check result is not hot data, the metadata server maintains the sleep mode by the check execution cycle (S920), and returns to the process of obtaining an entry (S903).

If the check result is hot data, the metadata server fetches an access count field of the hot data management table (S911), and changes the access count field based on the current time (S912). The metadata server determines whether the inspection period of the hot data is in units of one day (S930), and if the inspection period is in units of one day, the metadata server determines the value of day [0] to day [inspection period] in the access count field. The sum is added to the sum counter variable (S931). The metadata server obtains the current copy binding number of the file of the entry of the hot data management table (S960), and determines whether the value obtained by dividing the number of accesses by the current copy number is greater than the threshold value (S970). If it is determined that the threshold value is not large, the metadata server proceeds to step S904 of entry confirmation.

As a result of the determination in the determination process (S970), if the threshold value is larger, the metadata server recognizes the number of times of access of the file as hot data, replicates a copy of the file to a new data server (S971), and counts the number of file copies. Increase by 1 (S972). Subsequently, the metadata server updates an additional replication presence field of an entry of a hot data management table to true (S973).

If it is determined that the determination process (S930) is not a daily unit, the metadata server determines whether the inspection period of the hot data is a unit of time (S940). If the determination result is a time unit, the metadata server adds the sum value of the hour [0] to hour [inspection period] values of the access count field to the hit counter variable (S941), and then sets the current copy binding number. After the obtaining step (S960) is performed.

As a result of the determination in operation S940, if the inspection period is not in units of time, the metadata server determines whether the inspection period of hot data is in units of minutes (S950). If the determination result is minutes, the metadata server adds the sum value of min [0] to min [inspection period] of the access count field into a hit counter variable (S951), and then sets the current copy binding number. The process proceeds after the obtaining step (S960).

As a result of the determination in operation S950, if the inspection period is not in minutes, the metadata server outputs that an error has occurred in the inspection period value of the hot data (S980).

10 is a flowchart illustrating a process of counting additional copies of a file according to an embodiment of the present invention. Hereinafter, the metadata server is the same server as the metadata server 120 of FIG. 1.

Referring to FIG. 10, the metadata server obtains a check execution period from the hot data management table (S1001), and obtains a check period value from the hot data management table (S1002) to detect hot data. The metadata server retrieves from the hot data management table all entries whose recent access time is not in the inspection period from the current time (S1003), checks the imported entries one by one (S1004), and checks whether it is marked as hot data. (S1010). If the check result is not hot data, the metadata server maintains the sleep mode for the check execution cycle (S1050), and returns to the process of obtaining an entry (S1003).

If the check result is hot data, the metadata server determines whether an additional replication field is set in the hot data management table (S1020).

As a result of the determination in S1020, if the additional replication field is set, the metadata server reduces the number of replicas of the entry file by 1 (S1021), and then copies one of the replicas stored in the data server. Delete (S1022).

The metadata server determines whether the number of file replicas of the entry is equal to the predetermined reference number of replicas (S1030). If the number of copies is the same as a result of the determination, the metadata server initializes an additional copy field value of the entry (S1031), and proceeds to the entry verification process (S1004). Go back. If it is determined that the replica numbers are not the same, the process proceeds directly to the entry confirmation process (S1004).

If the additional replication field is not set as a result of the determination in S1020, the metadata server removes an entry from the hot data management table (S1040) and returns to an entry confirmation process (S1004).

That is, when the entry is taken in step S1003, entries that have not been accessed from the current time to the inspection period are imported. Determines that it is not hot data and performs the above-described process.

That is, through the embodiment of FIG. 10, storage data may be reduced by identifying data that was hot data in the past but not hot data and removing duplicate data.

As mentioned above, although the technical idea of this invention was described in detail with reference to the said preferable embodiment and an accompanying drawing, it should be noted that this is only an illustration and does not limit or limit this invention. Those skilled in the art of the present invention will be able to make various modifications and changes in the scope of the technical spirit of the present invention by acquiring the above-described description, the protection scope of the present invention by the description of the claims below It must be decided.

1 is a block diagram of an asymmetric storage system to which embodiments of the present invention are applied.

2 is a diagram illustrating a configuration of a data access number entry according to an embodiment of the present invention.

3 is a flowchart illustrating a method of processing data access times according to an embodiment of the present invention.

4 is a flowchart illustrating a method of changing a rank of data access times according to an embodiment of the present invention.

5 is a flowchart illustrating a method of processing a data access count according to an embodiment of the present invention.

6 is a diagram illustrating a configuration of a hot data management table according to an embodiment of the present invention.

7 is a flowchart illustrating a method of storing data access count information according to an embodiment of the present invention.

8 is a flowchart illustrating a method of updating an access count field according to an embodiment of the present invention.

9 is a flowchart illustrating a method of detecting and duplicating hot data according to an embodiment of the present invention.

10 is a flowchart illustrating a process of counting additional copies of a file according to an embodiment of the present invention.

Claims (20)

Monitoring the number of accesses to data stored by each data server of the asymmetric storage system to maintain the latest access count information for each data; Transmitting, by each data server, the access count information to a metadata server at predetermined intervals; Performing data duplication or deletion of the replica according to the determination of the metadata server by each data server. Hot data management method based on the number of accesses including. The method of claim 1, wherein maintaining the latest state Forming at least one data access number entry comprising a disk identifier field and a data identifier field for distinguishing data, an access count field for recording access counts, and a top list field; Preparing a hash table including a predetermined number of hash heads for quick identification of the access number entry; Managing access counts for each data using the data access number entry and the hash table; Hot data management method based on the number of access that includes. The method of claim 2, wherein the managing of the number of accesses for each data comprises: Receiving a data read or delete request, inserting a data identifier into a hash function to obtain a result value, and then checking whether there is a data access number entry with the data identifier in the hash head; Incrementing the access count field by one if a data access count entry exists and a read request; Creating a new data access count entry if the data access count entry does not exist and is a read request; If the data access number entry exists and is a delete request, removing the corresponding data access number entry Hot data management method based on the number of access that includes. The method of claim 3, And changing the rank of the data access number entry when the number of data accesses of the data access entry is increased. The method of claim 4, wherein changing the rank comprises: Searching for a data access number entry having a higher priority than the data access number entry with an increased data access number; Adjusting the rank of the data access count entry immediately after the searched data access number entry if found, and adjusting the data access number entry to the entry having the highest priority if not found. Hot data management method based on the number of access that includes. The method of claim 4, wherein changing the rank comprises: Identifying a previous data access number entry from a top list of data access number entries with an increased data access count; Making the previous data access count entry into the temporary entry; Setting the data access count entry to the next entry of the temporary entry if the data access count of the temporary entry is greater than or equal to the data access count of the data access count entry; If the data access number of the temporary entry is less than the data access number of the data access number entry, the previous entry on the top list of the temporary entry is updated with the temporary entry, and the data access number of the temporary entry is the data access number entry. Repeating this until it is greater than or equal to the number of times data is accessed Hot data management method based on the number of accesses including. 4. The method of claim 3, wherein generating the new data access number entry Creating a data structure comprising a disk identifier field and a data identifier field, an access count field for recording access counts, and a top list field for concatenating each data access count entry; Putting a disc identifier and a data identifier for the data into the disc identifier field and the data identifier field; Initializing the data access count field to '1'; Putting a hash result value of the data identifier into a hash head of one of the hash tables; Determining the top list field value to have a lowest priority Hot data management method based on the number of accesses including. The method of claim 1, wherein the transmitting step Transmitting information on the number of times of access to the data by the maximum number of transmissions set in advance in order of increasing number of times of access to the data; Initializing the data access number entry; Hot data management method based on the number of access that includes. Configuring a hot data management table including an access count field for each data; Collecting access count information for data stored by each data server from one or more data servers; Updating the hot data management table according to the access count information; Determining the hot data by checking the hot data management table at predetermined intervals; Duplicating the data file determined to be hot data to a new data server; Deleting a replica of the data file that is no longer hot data Hot data management method based on the number of accesses including. The method of claim 9, wherein updating the hot data management table Obtaining a file identifier from a disk identifier and a data identifier included in the received access count information; If an entry with the file identifier exists in the hot data management table, updating the entry based on the access count information; Adding and initializing a new entry if there is no entry with the file identifier in the hot data management table Hot data management method based on the number of access that includes. The method of claim 10, wherein updating the corresponding entry comprises: Obtaining the elapsed day, elapsed time, or elapsed time from the last update time to the present time point, Min [0] to min [63], hour [0] to hour [23], or day [0] to day [364] arrays constituting the number of access fields, respectively, by the elapsed minutes, elapsed days, or elapsed days. Shifting to the right, Increasing the min [0] value, hour [0] value, or day [0] value by the data access count value; Updating the latest access time to the current time Hot data management method based on the number of access that includes. The method of claim 10, wherein adding and initializing the new entry comprises: Initializing min [0] to min [63], hour [0] to hour [23], and day [0] to day [364] of the access number field of the new entry to '0'; Setting last_min, last_hour and last_day in the number of access fields of the new entry as the current time; Putting the new entry in the hot data management table Hot data management method based on the number of accesses including. The method of claim 9, wherein determining whether the hot data is performed Checking the inspection run cycle; Calculating the number of accesses during the period in the inspection execution cycle based on the current time, for the entry updated in the inspection execution cycle based on the current time among the entries in the hot data management table; Ascertaining the number of copies of the data file corresponding to the entry; Checking whether the number of accesses divided by the number of replicas exceeds a predetermined threshold; If exceeding, determining the data corresponding to the entry as hot data Hot data management method based on the number of access that includes. The method of claim 13, wherein said calculating Obtaining the number of accesses by adding day [0] to day [inspection period] values of the access number field if the inspection execution cycle is a unit of one day; Obtaining the number of accesses by adding hour [0] to hour [inspection period] values of the access number field if the inspection execution period is a unit of time; Calculating the number of accesses by adding min [0] to min [inspection period] values of the access number field if the inspection execution period is in minutes; Hot data management method based on the number of access that includes. 10. The method of claim 9, wherein said replicating is Duplicating the data determined to be hot data to a new data server; Increasing the number of file replicas by one; Setting an additional copy existence field of the entry of the hot data management table to true Hot data management method based on the number of access that includes. The method of claim 9, wherein deleting the replica is Identifying entries in the hot data management table that were hot data in the past but are not currently hot data; Reducing the number of copies of the file corresponding to the identified entry by one; And Steps to delete one of the existing replicas Hot data management method based on the number of access that includes. The method of claim 16, Initializing additional duplicate field values of the entry if the number of replicas of the file is equal to a predetermined reference replica number; Hot data management method based on the number of accesses further including. The method of claim 16, If an additional duplicate field is not set in the checked entry, removing the checked entry from the hot data management table Data management method further comprising. The method of claim 16, wherein the identifying step Checking a value of a recent access time field of each entry in the hot data management table; Selecting entries for which the identified last access time does not fall within the inspection period from the current time; Hot data management method based on the number of access that includes. Maintaining, by a plurality of data servers, the number of times data is accessed for data stored and managed by the plurality of data servers; Transmitting, by the plurality of data servers, the number of data accesses to the stored data to the management server at each predetermined period; Collecting and storing the number of data accesses transmitted by the management server; The management server recognizes, as a corresponding hot data, data for which the number of times of access to each data server exceeds a preset threshold for each predetermined period and for each period, and the data of the hot file is one or more data servers of the plurality of data. Additional cloning steps Hot data management method comprising a.

Images